Matthias Raunhardt- Generation and spectroscopy of atoms and molecules in metastable states

8/3/2019 Matthias Raunhardt- Generation and spectroscopy of atoms and molecules in metastable states

1/203

Diss. ETH No. 18260

Generation and spectroscopy of

atoms and molecules in metastable

states

A dissertation submitted to

ETH ZURICH

for the degree of

Doctor of Sciences

presented by

Matthias Raunhardt

Dipl. Phys. ETH

born February 2, 1979citizen of Zurich

Accepted on the recommendation of:

Prof. Dr. F. Merkt, examiner

Prof. Dr. A. Wokaun, co-examiner

2009


2/203


3/203

Meinen Eltern und meinem Bruder


4/203


5/203

Ins Innre der Natur

Dringt kein ersaner Gei.Gluselig, wem sie nur

Die aure Sale wei.

Albret von Haller


6/203


7/203

Abstract

A new spectrometer for the study of metastable states of atoms and molecules by

high-resolution photoionization spectroscopy, pulsed-field-ionization zero-kinetic-energy

(PFI-ZEKE) photoelectron spectroscopy, and mass spectrometry has been designed, built

and tested. The apparatus consists of 1) a versatile pulsed laser system that enables in-

vestigations by single-photon excitation using coherent ultraviolet and vacuum-ultraviolet

(VUV) radiation, and by multiphoton excitation, 2) a photoionization/photoelectron spec-

trometer especially designed for experiments on very high Rydberg states such as those

probed by PFI-ZEKE photoelectron spectroscopy, 3) a probe gas source chamber in whichmetastable states of atoms and molecules can be generated using pulsed discharges, pho-

tolysis, and single- and multiphoton optical excitation of forbidden transitions, and 4) a

source of narrow-bandwidth tunable millimeter-wave radiation for high-resolution spec-

troscopy of high Rydberg states. The spectrometer has been used to study the threshold

ionization dynamics of ND3, the photolysis of ammonia cluster and the threshold ion-

ization of ammoniated ammonium radical clusters, the hyperfine structure of long-lived

high Rydberg states of 129Xe and 131Xe, the formation of metastable He and He2 in su-

personic expansions, the structure of He+2

, and the rotational autoionization dynamics of

high Rydberg states of He2.

The PFI-ZEKE photoelectron spectrum of ND3 was recorded following single-photon

VUV excitation from the X 1A1 ground state at a resolution and sensitivity sufficient to

observe the rotational structure of the first three out-of-plane bending vibrational levels

of ND+3 . The analysis of the photoelectron spectrum enabled the determination of the

first adiabatic ionization energy of ND3 (82 261.7 1.5 cm1) and of a set of molecu-lar constants describing the rovibrational energy level structure of the X

+ 2A2 ground

electronic state of ND+3 up to 1400cm1 of internal energy. The intensity distribution of

the photoelectron spectrum was analyzed in terms of rovibronic photoionization selection

rules, a simple orbital ionization model, and rovibronic channel interactions. The dom-

inant ionization channels were found to be associated with photoelectron partial waves

of even orbital angular momentum quantum number . Intensity perturbations resulting

from vibrational channel interactions were identified and found to significantly modify

the photoelectron partial wave composition near the position of strong autoionization

resonances in the photoionization spectrum.

The formation of metastable undeuterated and perdeuterated ammonium radicals and

ammoniated ammonium radicals by UV photolysis of ammonia clusters in supersonic

expansions was studied by photoionization mass spectrometry and PFI-ZEKE photo-

vii


8/203

viii Abstract

electron spectroscopy. The 3s 2A1 state of the perdeuterated ammonium radical was

found to have a lifetime of about 20 s, more than four orders of magnitude longer

than the lifetime of the 3s 2A1 state of the NH4 radical. Very pronounced isotopic ef-

fects were also observed in the photolysis of ammonia trimers ((NH3)3 and (ND3)3) at

204 nm: Whereas the metastable NH3NH4 radical was identified as the dominant prod-

uct in the case of the undeuterated clusters, ND2ND4 represented the main product in

the case of the perdeuterated clusters. The PFI-ZEKE photoelectron and photoioniza-

tion spectra enabled the determination of the adiabatic ionization energy of NH3NH4

(31 310(25) cm1) and ND3ND4 (31 220(25) cm1) and of a very pronounced ionization

threshold at 30 780 cm1 in ND2ND4. The PFI-ZEKE photoelectron spectrum also pro-

vided information on low-lying vibrational levels of the N2H+7 cation. Two low-lying

vibrational levels were observed at 345 cm1 and 690cm1 above the origin that were

tentatively assigned to the fundamental and first overtone of the symmetric stretching

vibrational mode involving the two NH3 molecules on both side of the proton on the basis

of a comparison with the IR spectrum of N2H+7 and ab initio calculations [Yang et al., J.

Chem. Phys. 129, 224302 (2008), Garca-Fernandez et al., J. Chem. Phys. 129, 124313

(2008)].

Millimeter-wave spectra of transitions between metastable high ns, np, nd, and nf

Rydberg states of Xe were recorded at a resolution sufficiently high ( 500 kHz) to observethe hyperfine structure in the spectra of 129Xe and 131Xe. These spectra were analyzed

using multichannel quantum defect theory (MQDT) and led to the first determination ofthe hyperfine structure of the 2P3/2 ground state of

129Xe+ and 131Xe+. A set of MQDT

parameters was derived for the ns and nd Rydberg states that accurately describes the

photoionization dynamics of Xe and all spectroscopic data obtained up to date on neutral

Xe.

A new method to generate the 1s2s 3S1 metastable state of He and the a3+u metastable

state of He2 in supersonic expansions has been developed. Photoionization and PFI-ZEKE

photoelectron spectra of 4He2 and3He2 have been recorded and analyzed and a set of ac-

curate molecular constants for the lowest three vibrational levels of the ground electronicstate of He+2 has been derived. The use of Rydberg-state-resolved threshold-ionization

spectroscopy in combination with photoionization spectroscopy and multichannel quan-

tum defect theory has provided a detailed description of the rotational autoionization

dynamics of the triplet np Rydberg states of 4He2 has been derived. Series of metastable

Rydberg states embedded in the ionization continua were observed, and the origin of their

metastability explained by a simple MQDT model. The application of weak electric fields

was found to induce rotational autoionization of these long-lived Rydberg states, inducing

changes of the rotational angular momentum quantum number of the 4He+2

ion core by

up to six units.


9/203

Zusammenfassung

Fur die Untersuchung von Atomen und Molekulen in metastabilen Zustanden mittels

hochauflosender Photoionisationsspektroskopie, pulsed-field-ionization zero-kinetic energy

(PFI-ZEKE) Photoelektronenspektroskopie und Massenspektrometrie wurde ein neues

Spektrometer entwickelt, gebaut und getestet. Das Spektrometer besteht aus 1) einem

gepulsten Lasersystem, das Untersuchungen von Einphotonen- und Mehrphotonenan-

regungen mittels koharenter ultravioletter und vakuumultravioletter (VUV) Strahlung

ermoglichte, 2) einem Photoionisations-/Photoelektronenspektrometer, das speziell fur

Experimente an sehr hohen Rydbergzustanden entwickelt wurde, 3) einer Probengaskam-mer, die zur Erzeugung von metastabilen Zustanden von Atomen und Molekulen mittels

gepulster elektrischer Entladungen, Photolyse und Einphotonen- und Mehrphotonenan-

regungen verbotener Ubergange genutzt wurde, und 4) einer schmalbandigen, abstimm-

baren Millimeterwellenquelle, mit der hohe Rydberzustande bei sehr hoher Auflosung

untersucht wurden. Das Spektrometer wurde fur die Untersuchung der Schwellenionisa-

tionsdynamik von ND3, der Photolyse von NH3-Aggregaten, der Ionisation von NH3NH4-

Radikalaggregaten, der Hyperfeinstruktur von langlebigen, hohen Rydbergzustanden von129Xe und 131Xe, der Erzeugung von metastabilem He und He

2in Uberschallstrahlexpan-

sionen, der Struktur von He+2 , sowie der Rotationsautoionisationsdynamik hoher Ryd-

bergzustande von He2 verwendet.

Das PFI-ZEKE Photoelektronenspektrum von ND3 wurde durch Einphotonenanre-

gung aus dem X 1A1 Grundzustand gemessen, wobei die Auflosung und Empfindlichkeit es

erlaubten, die Rotationsstruktur der ersten drei out-of-plane Biegeschwingungsniveaus

von ND+3 zu beobachten. Die Auswertung des Photoelektronenspektrums ermoglichte die

Bestimmung der ersten adiabatischen Ionisationsenergie von ND3 (82 2261.71.5 cm1)und eines Satzes molekularer Konstanten, welche die rovibrationelle Energieniveaustruk-

tur des X+ 2A2 elektronischen Grundzustandes von ND

+3 bis zu einer internen Energie von

1400cm1 beschreibt. Die Intensitatsverteilung des Photoelektronenspektrums wurde mit

Hilfe von rovibronischen Photoionisationsauswahlregeln, einem einfachen Orbitalionisa-

tionsmodell und rovibronischen Kanalwechselwirkungen analysiert. Aus den beobachteten

Ubergangen konnte geschlossen werden, dass die dominierenden Ionisationskanale Pho-

toelektronenpartialwellen gerader Bahndrehimpulsquantenzahl entsprechen. Dabei wur-

de beobachtet, dass Intensitatsstorungen, welche von vibrationellen Kanalwechselwirkun-

gen stammen, die Zusammensetzung der Photoelektronenpartialwellen in der Nahe von

starken Autoionisationsresonanzen im Photoionisationsspektrum deutlich veranderten.

Durch Photoionisationsmassenspektrometrie und PFI-ZEKE Photoelektronenspektro-

ix


10/203

x Zusammenfassung

skopie wurde die Bildung von metastabilen undeuterierten und komplett deuterierten

Ammoniumradikalen sowie von metastabilen Aggregaten aus Ammoniak und Ammonium-

radikalen durch UV Photolyse von Ammoniakaggregaten in einem Uberschallstrahl unter-

sucht. Es wurde eine Lebenszeit von ca. 20 s fur den 3s 2A1 Zustand von ND4 ermittelt,

die mehr als vier Grossenordnungen langer ist als die Lebenszeit des entsprechenden 3s2A1 Zustandes des NH4-Radikals. Auch bei der Photolyse des Ammoniaktrimers ((NH3)3

and (ND3)3) bei 204 nm wurde ein ausgepragter Isotopeneffekt beobachtet: wahrend im

Falle der undeuterierten Aggregate metastabile NH3NH4 Molekule als vorherrschendes

Produkt identifiziert wurden, stellte im Falle der volldeuterierten Aggregate ND2ND4

das Hauptprodukt dar. Die PFI-ZEKE Photoelektronen- und Photoionisationsspektren

ermoglichten die Bestimmung der adiabatischen Ionisierungsenergien von NH3NH4

(31 310(25) cm1), ND3ND4 (31 220(25) cm1) und einer ausgepragten Ionisationsschwelle

bei 30 780 cm1 in ND2ND4. Das PFI-ZEKE Photoelektonenspektrum enthielt auch In-

formationen uber tiefliegende Vibrationszustande des N2H+7 Kations. Zwei tiefliegende

Vibrationszustande wurden bei 345cm1 und 690cm1 oberhalb der Ursprungsbande

beobachtet. Auf Grund eines Vergleiches mit dem IR Spektrum von N2H+7 und mit ab

initio Rechnungen wird eine Zuordnung dieser zwei Zustande zur Grundschwingung und

zum ersten Oberton der Streckschwingungsmode der beiden NH3 Molekule auf beiden

Seiten des Protons hin vorgeschlagen, die allerdings noch bestatigt werde muss [Yang et

al., J. Chem. Phys. 129, 224302 (2008), Garca-Fernandez et al., J. Chem. Phys. 129,

124313 (2008)].

Zusatzlich wurden Millimeterwellenspektren von Ubergangen zwischen metastabilen

hohen ns, nd, np and nf Rydbergzustanden von Xenon gemessen. Die Auflosung dieser

Spektren war hoch genug ( 500 kHz), um die Hyperfeinstruktur in den Spektren von129Xe und 131Xe zu beobachten. Die Analyse der Spektren mittels der Vielkanal-Quanten-

defekt-Theorie (MQDT) fuhrte zur ersten Bestimmung der Hyperfeinstruktur des 2P3/2

Grundzustandes von 129Xe+ und 131Xe+. Dabei wurde ein Satz von MQDT Parametern

fur die ns und nd Rydbergzustande ermittelt, der die Photoionisationsdynamik von Xenon

und alle spektroskopischen Information, welche bis heute erhalten wurden, beschreibt.

Durch eine neuentwickelte Methode konnte der 1s2s 3S1 metastabile Zustand von He

und der a 3+u metastabile Zustand von He2 in einer Uberschallstrahlexpansion erzeugt

werden. Die Analyse gemessener Photoionisations- und PFI-ZEKE Photoelektronen-

spektren von 4He2 und3He2 ermoglichte die Bestimmung eines Satzes von molekularen

Konstanten fur die drei tiefsten Vibrationsniveaus des elektronischen Grundzustandes

von He+2 . Die Kombination von Rydbergzustandsaufgeloster-Schwellenionisationsspektro-

skopie (Rydberg-state-resolved threshold-ionization spectroscopy), Photoionisationsspek-

troskopie und MQDT Analyse fuhrte zu einer detaillierten Beschreibung der Autoionisa-

tionsdynamik der S = 1 np Rydbergzustande von 4He2. Es wurden Serien metastabiler


11/203

Zusammenfassung xi

Rydbergzustande oberhalb der tiefsten Ionisationsschwelle beobachtet und der Ursprung

ihrer Metastabilitat durch ein einfaches MQDT Modell erklart. Die rotationelle Autoioni-

sation dieser metastabilen Zustande konnte durch das Anlegen von schwachen elektrischen

Feldern induziert werden, die zu einer Anderung der Rotationsquantenzahl im 4He+2 Io-

nenrumpf von bis zu sechs Einheiten in fuhrte.


12/203

xii


13/203

Contents

Abstract vii

Zusammenfassung ix

1 Introduction 1

1.1 Metastable states of atoms and molecules . . . . . . . . . . . . . . . . . . . 1

1.2 Outline of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3 Rydberg states a brief introduction . . . . . . . . . . . . . . . . . . . . . 4

2 Experimental setup and methods 7

2.1 Spectrometer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 The laser systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

2.2.1 Experiments using VUV radiation . . . . . . . . . . . . . . . . . . . 112.2.2 Experiments using single-photon excitation with UV radiation . . . 12

2.2.3 Experiments using multiphoton excitation . . . . . . . . . . . . . . 13

2.2.4 Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.3 Millimeter-wave source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.4 Probe gas samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.4.1 Studies of molecular cations using seeded supersonic expansion . . . 16

2.4.2 Discharge source of metastable He and He2 in supersonic expansions 17

2.4.3 Optical preparation of metastable states . . . . . . . . . . . . . . . 202.4.4 Photolysis source of metastable radicals in supersonic expansions . 22

2.5 Spectroscopic methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

2.5.1 Photoionization spectroscopy . . . . . . . . . . . . . . . . . . . . . 23

2.5.2 PFI-ZEKE photoelectron spectroscopy . . . . . . . . . . . . . . . . 26

2.5.3 Millimeter-wave spectroscopy of high Rydberg states . . . . . . . . 28

3 The PFI-ZEKE photoelectron spectrum of ND3 31

3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.2 Rovibronic photoionization selection rules of ND3 . . . . . . . . . . . . . . 34

3.3 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36

xiii


14/203

xiv Contents

3.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

3.5 Discussion and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

4 Threshold photoionization of ammoniated ammonium clusters 45

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.2 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

4.3 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.3.1 Formation and threshold ionization of the deuterated ammonium

radical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.3.2 Formation and threshold ionization of the ammoniated ammonium

dimers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

4.3.2.1 Experiments on (NH3)3 clusters . . . . . . . . . . . . . . . 50

4.3.2.2 Experiments on (ND3)3 clusters . . . . . . . . . . . . . . . 53

4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

5 Spectroscopy and MQDT analysis of high Rydberg states of Xe 57

5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5.2 Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

5.3 MQDT calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.5 Discussion and conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

6 PFI-ZEKE photoelectron spectroscopy of metastable He and He2 83

6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83

6.2 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

6.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

6.3.1 Performance of the pulsed discharge . . . . . . . . . . . . . . . . . . 86

6.3.2 The photoionization and PFI-ZEKE photoelectron spectra of He . 89

6.3.3 The PFI-ZEKE spectra of He

2 and their assignment . . . . . . . . . 916.3.4 The photoionization spectrum He2 and its analysis by MQDT . . . 97

6.3.5 Probing the autoionization dynamics of He2 Rydberg states by

Rydberg-state-resolved threshold-ionization . . . . . . . . . . . . . 102

6.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

6.4.1 Energy level structure of the He+2 cation . . . . . . . . . . . . . . . 106

6.4.2 The field ionization and rotational autoionization dynamics of Ryd-

berg states of He2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

6.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

7 Conclusions 117


15/203

Contents xv

A Physical constants 119

B Spectral positions of the observed Rydberg levels of He2 121

List of figures 143

List of tables 153

Bibliography 157

Curriculum Vitae 183

Publications 185

Danksagung 187


16/203


17/203


18/203

2 Introduction

levels of the ground electronic state of molecules, result from the conservation of nuclear

spin symmetry and/or parity and/or the absence of a permanent dipole moment.

For a given state of an atom or a molecule to be metastable it is necessary that all

radiative and nonradiative processes such as predissociation, autoionization and internal

conversions be suppressed. This suppression inevitably occurs in Rydberg states of atoms

and molecules of high principal quantum number n belonging to series converging to the

lowest-lying quantum state of a positively charged ion. Indeed, when n tends towards

infinity, the Rydberg electron wavefunction has a vanishingly small amplitude close to

the ion core, and the motion of the electron gets entirely decoupled from that of the ion

core; under these conditions, neither radiative nor nonradiative decay processes can take

place.

Atoms and molecules in metastable states are important carriers of energy in chemical

reactions, the a 1g and b 1+g states of O2 and the lowest3P state of Hg being among the

most prominent examples. Metastability of an isolated atom or molecule is thus closely

linked to, and almost inevitably results in a high reactivity.

Metastable states of atoms and molecules are also attractive systems for studies by

high-resolution spectroscopy. Their long lifetimes allow long measurement times and,

by virtue of the energy-time uncertainty principle, high spectral resolution. Because

their electronic or nuclear configurations usually differ from those of the ground state,

metastable states offer the possibility to study, by spectroscopy, states that are not ac-

cessible from the ground state. Finally, metastability is closely connected to unusual

properties of potential energy surfaces, or unusual nuclear or electronic configurations,

and consequently metastable atoms and molecules are per se interesting objects of scien-

tific investigations.

This dissertation is devoted to spectroscopic studies of atoms and molecules in metasta-

ble states in the gas phase. Three aspects have been at the foreground: 1) The im-

plementation of efficient methods to generate dense samples of metastable atoms and

molecules, 2) The use or development of very sensitive spectroscopic methods to study

these metastable atoms and molecules, and 3) The exploration of as diverse forms ofmetastability as possible.

1.2 Outline of the thesis

This introductionary chapter will give in the next subsection a brief introduction to Ryd-

berg states that have been central to this dissertation not only as an important class of

metastable states, but also because the methods used to detect metastable atoms and

molecules such as pulsed-field-ionization zero-kinetic-energy (PFI-ZEKE) photoelectron

spectroscopy and Rydberg-state-resolved threshold-ionization (RSR-TI) spectroscopy rely


19/203

Outline of the thesis 3

on the unusual physical properties of these states.

Chapter 2 focuses on the description of the apparatus that has been built and the

methods that have been implemented to generate dense samples of metastable molecules

and to study them. Metastable states of He, He2, Xe, CO, ND4, ND2-ND4, N2H7, N2D7were populated using specific methods including optical excitation from the ground state

by driving forbidden transitions with intense laser pulses, photolysis of suitable precur-

sors in supersonic expansions, multiphoton excitation from the ground state, and electric

discharges. They have been studied by photoionization spectroscopy, photoelectron spec-

troscopy (mostly PFI-ZEKE and RSR-TI spectroscopy), millimeter-wave spectroscopy

and time-of-flight mass spectrometry.

Chapter 3 is devoted to a study of the photoionization and PFI-ZEKE photoelectron

spectrum the ND+3

X+ 2A

2 ND3 X

1A1

transition. Though neither the ground electronic

state of ND3 nor that of ND+3 classify as metastable states, the high Rydberg states of

ND3 used to record the photoelectron spectra are metastable. Particular emphasis was

placed on the characterization of the photoionization dynamics, and of the high Rydberg

states and ionization channels accessible from the ground neutral state.

Chapter 4 investigates the metastability with respect to dissociation of the ammo-

nium radical and ammoniated ammonium radicals and their cations. Metastability in

these molecules is strongly influenced by isotopic substitution, and the very pronounced

differences in the behavior of undeuterated and perdeuterated samples proved very useful

to characterize the origin of the metastability of these systems.

The fifth chapter summarizes the study of long-lived s, p, d and f high Rydberg states of

atomic xenon by high-resolution millimeter-wave spectroscopy and multichannel quantum

defect theory, a collision theory. The metastability of high Rydberg states originates from

the decoupling of the motions of the Rydberg electron and of the ionic core. One of the

main goals of this study was to resolve the hyperfine structure in high Rydberg states of129Xe and 131Xe to see whether the interaction of the Rydberg electron with the nuclear

spin of the ion core can be observed in these metastable states despite the fact that the

electron and ion core motions are decoupled and to study how nuclear spin influence the

photoionization dynamics.

Chapter 6 presents the results of an investigation of the photoionization and photo-

electron spectra of helium atoms in the metastable 1s2s 3S1 He state and of He2 molecules

in the a 3+u metastable state. The challenge was to derive a method of generating the

metastable species in a supersonic expansion under conditions compatible with the record-

ing of photoionization and photoelectron spectra. This study resulted in the spectroscopic

characterization of the ground state of the He+2 cation and in an investigation of the ro-

tational autoionization dynamics of high Rydberg states of He2. The emphasis was on

the search, identification and characterization of states that are metastable with respect


20/203

4 Introduction

to ionization. Many such states could be identified. However, weak electric fields were

found to induce rapid ionization of the metastable states.

The thesis ends in Chapter 7 with a brief summary and an outlook.

1.3 Rydberg states a brief introduction

Rydberg states represent an important class of metastable states in atoms and molecules.

Their metastability is at the heart of important spectroscopic techniques such as PFI-

ZEKE photoelectron spectroscopy [6,7], mass-analyzed-threshold-ionization (MATI) spec-

troscopy [8], photoinduced Rydberg ionization (PIRI) spectroscopy [9,10], and Rydberg-

tagging photofragment translational spectroscopy [1113].

Rydberg states are electronically excited states of neutral or positively charged atomsand molecules the spectral positions n, of which can be described by Rydbergs formula

n, =EIhc

RM(n )2

, (1.1)

in which EI, the ionization energy, represents the convergence limit of the infinite series of

levels with principal quantum number n (n = 1, 2, 3,...) and orbital angular momentum

quantum number . RM is the mass-dependent Rydberg constant which can be expressed

in terms of the Rydberg constant for an atom of infinite mass R as (M+

is the mass ofthe positively charged ionic core)

RM = RM+

M+ + me. (1.2)

In Eq. (1.1) represents the quantum defect which is zero for all series of atomic hydrogen

and one-electron atoms, and for all nonpenetrating series (i.e., series with typically larger

than 3) but nonzero for core-penetrating series (typically 3) of atoms and moleculeshaving more than one electron.

Most physical properties of Rydberg states scale as integer power of the principal

quantum number, and Table 1.1 provides an overview of the n-scaling laws of selected

properties with numerical values at n = 1, 10 and 100. The n-scaling law can easily

derived by considering the wavefunctions |nm of the hydrogen atom. Of particularsignificance in the realm of the present dissertation are the following properties: 1) The

lifetimes of Rydberg states of a given series increase with n3. Therefore high-n Rydberg

states are metastable. The n3 scaling of the lifetime can be understood from the fact

that the decay of a Rydberg state, be it radiative, by autoionization or predissociation,

is proportional to the probability of finding the electron in close vicinity of the ion core,

i.e., to the square of the Rydberg electron wavefunction in this region. 2) The spacing


21/203

Rydberg states a brief introduction 5

Table 1.1: Properties of Rydberg states. Values are calculated for the hydrogen atom if not mentioned

otherwise

Value atProperty n-scaling n = 1 n = 10 n = 100

Radius / A a0n2

0.53 53 5300Binding energy / cm1 Rn2 109667 1097 0.1Lifetimea / s n3Orbital period / s 1n

3 1.5 1016 1.5 1013 1.5 1010Spacing between neighboring states / cm1 2Rn3 82 000 220 0.22Polarizabilityb / MHz/(V/cm)2 n7 2.1 108 0.21 2.1 107Threshold ionization field / (V/cm) n4 108 3.2 104 3.2

aThe numerical values strongly depend on the series and on the atom or moleculebValues for the nd series of Na extrapolated from the measured value of the 10d state [14]

between adjacent members of a given series scales as 2Rn3 as can easily be derived by taking

the derivative of Eq. (1.1). High-resolution spectroscopic methods are thus mandatory if

one wants to resolve individual members of a Rydberg series at high n. 3) The electric

field that is necessary to ionize a Rydberg state scales as n4 and decreases very rapidly

with n. This property renders the detection of high Rydberg states easy and efficient. A

dc electric field lowers the ionization limit by EI

EI

hc cm

1 = 6.12E

V / c m

(1.3)

and a pulsed electric field causes ionization of Rydberg states located within

EIhc cm1

= C

E

V / c m(1.4)

of the ionization limit, where C can take values ranging from 3 to 6.12 depending on the

field ionization mechanism [1416]. 4) The polarizability of Rydberg states scales as n7.

In order to study Rydberg states at high n values it is thus imperative to minimize stray

electric fields.

The rapid variation of the physical properties of Rydberg states with n implies that

the properties of high Rydberg states are exaggerated which can be exploited in a variety

of applications such as the spectroscopic methods mentioned at the beginning of this

subsection, the measurement of stray fields and ion concentrations [17], the electrostatic

acceleration and deceleration of beams of Rydberg atoms and molecules [18,19], and the

electrostatic trapping of translationally cold samples of Rydberg atoms [20].


22/203

6 Introduction


23/203

Chapter 2

Experimental setup and methods

This chapter describes the spectrometer and spectroscopic methods that have been used

in the realm of the present dissertation. The laser system, the vacuum chambers, and

the source of metastable molecules have been specially designed, constructed and tested

to carry out spectroscopic measurements on unstable, reactive or metastable molecules

by exciting them to high Rydberg states close to their ionization thresholds. These

species include molecular cations, electronically excited long-lived atoms and molecules

and radicals. Several different photoexcitation schemes have been employed, including

single-photon excitation or ionization with narrow-band VUV radiation (study of ND+3 in

Chapter 3) or UV radiation (studies of He and He 2 and He+2 in Chapter 6 and of ND4 and

ammonium clusters in Chapter 4), and resonance-enhanced multiphoton excitation ion-

ization (study of Xe Rydberg states in Chapter 5). This chapter begins in Section 2.1 with

a description of the photoionization and photoelectron spectrometer, which was common

to all experiments. This section is followed by a description of the laser and millimeter-

wave radiation sources in Sections 2.2 and 2.3, respectively, and of the different methods

used to generate metastable molecules in Section 2.4. Finally, the last section provides an

introduction to the three spectroscopic methods employed: photoionization spectroscopy,

pulsed-field-ionization zero-kinetic-energy (PFI-ZEKE) photoelectron spectroscopy and

millimeter-wave spectroscopy.

2.1 Spectrometer

The spectrometer developed and used in the realm of this dissertation is depicted schemat-

ically in Fig. 2.1. It consists of a laser system, a probe gas chamber and a photoexci-

tation/photoionization chamber. The different configurations of the laser system used

for the different experiments summarized in this thesis will be described in more de-

tail in Section 2.2. This section presents the probe gas chamber and the photoexcita-

tion/photoionization chamber that were common to all experiments. An enlarged view

7


24/203

8 Experimental setup and methods

of these chambers is shown in Fig. 2.2.

Figure 2.1: Schematic view of the experimental setup for experiments on metastable molecules, radicals

and ions. The lasers with wave number 1 and 2 were used to generate VUV radiation of wave number

VUV. Alternatively they could be used to induce two-color multiphoton excitation processes, in which

case the four-wave mixing chamber and the monochromator were not used. The vacuum chambers where

the photoexcitation and photoionization experiments were carried out are displayed on the right and the

detection and measurement systems at the bottom of the figure.

Figure 2.2: Schematic view of the photoelectron/photoion time-of-flight mass spectrometer. Depending

on the sign of the voltages applied to the set of six extraction plates (middle of the figure) and the

microchannel plate detectors, either ions or electrons can be detected using the same spectrometer. The

electron or ion time-of-flight tube and the photoexcitation region are surrounded by a magnetic shield

consisting of two concentric mumetal cylinders.

Photoionization and photoelectron spectroscopic experiments must be carried out un-

der high vacuum. Pulsed skimmed supersonic expansions in which the species of interest

is diluted in a carrier gas (typically a noble gas) were chosen as gas samples for the

following reasons: 1) They offer a collision-free environment which is a prerequisite in

experiments including reactive species. 2) They enable one to easily maintain a very

low pressure (typically less than 106 mbar) in the regions of the spectrometer where the


25/203

Spectrometer 9

charged particles are detected higher pressures tend to lead to undesirable electric dis-

charges in the vicinity of electron and ion detectors which are kept at high voltages. 3)

The efficient cooling of vibrational and rotational degrees of freedom in supersonic beams

greatly reduce the number of populated quantum states and permit the optimization of

the population of the lowest rotational states of the molecular species. 4) The use of a

skimmer to separate the probe gas chamber from the photoionization chamber enables

one to carry out the experiments on highly directional atomic or molecular beams and

thus to completely suppress Doppler broadening. It also offers the possibility of exploit-

ing differential pumping between the probe gas chamber, which is exposed to a high gas

load (pressure of 104 mbar were typical for the experiments described in this thesis),

and the photoionization region, which is exposed to a much lower gas load and in which

pressures below 106 mbar are mandatory as explained above. To reach these pressures,

the probe gas chamber and the photoionization chamber were pumped by turbomolecular

pumps of 520ls1 (Pfeiffer TMH 521) and 260 l s1 (Pfeiffer TMH 261) pumping power,

respectively.

The probe gas samples were introduced into the gas chamber using a pulsed solenoid

valve (General Valve, Series 9, with IOTA One driver; nozzle orifice diameter of 0.5 mm).

The repetition rate of the nozzle opening (25 Hz) matched the repetition rate of the

laser system. The time difference between the opening times of the nozzle and the laser

pulses was adjusted using a high-precision delay generator (Stanford Digital Delay/Pulse

Generator Model DG535) to optimize the signal of the desired species and/or to minimize

the rotational temperature of the sample. For selected experiments, a pulsed discharge

setup or a quartz capillary crossed by a photolysis laser beam were mounted at the exit

of the pulsed valve in order to generate reactive or metastable species just before the

supersonic expansion (see Section 2.4).

Because most experiments reported in this thesis involved the photoexcitation to high

Rydberg states and their subsequent field-ionization by pulsed electric fields, great care

was taken in the design of the spectrometer to minimize stray electric and magnetic fields

and ion concentrations to which high Rydberg states are extremely sensitive [14, 15, 21]:

1) To minimize magnetic fields in the photoexcitation region the extraction plates and

the attached flight tube were surrounded by two cylindrical mumetal tubes in a geometry

ensuring an attenuation of magnetic fields by at least three orders of magnitude. 2) The

metallic plates used to apply the pulsed electric fields were carefully polished and made

of demagnetized stainless steel. Their geometric arrangement (stack of six equidistant

cylindrical parallel plates of total length of 7.8 cm length, each plate with a large central

opening of 1.8 cm) was designed so that the minimal distance of any metallic surface

from the photoexcitation region was at least 1 cm. In this way, the effects of stray fields

arising at defects of the metal surfaces could be minimized and the applied electric fields


26/203


were very homogeneous across the photoexcitation region. 3) The cylindrical plates were

carefully coupled by resistors and capacitors to ensure excellent impedance matching

with the voltage pulse supplies. Large voltage pulses up to 6000 V could be applied

with rise times of less than 50 ns, and ripples in the high voltages following the voltage

rise time could be suppressed [22]. 4) The cylindrical symmetry of the spectrometer

ensured minimal stray fields perpendicular to the symmetry axis, and the electrode stack

permitted the accurate compensation of stray electric fields along the symmetry axis.

5) Finally, the gas inlet and pumping systems with a skimmer separating the probe gas

chamber from the photoexcitation region, the optical access, and the laser intensities were

carefully optimized to avoid undesirable effects arising from interactions of the Rydberg

states with atoms and molecules in the background gas and with ions generated in the

photoexcitaion volume.

The necessity for these measures had been recognized prior to this thesis [17,22,23] but

the photoexcitation region designed and built in the realm of this dissertation represented

the first attempt to consider, in the design of a spectrometer in our group, all factors known

to be important in studies of high Rydberg states by high-resolution spectroscopy.

The charged particles produced by photoionization or by pulsed-field-ionization fol-

lowing photoexcitation to a long-lived Rydberg state and extracted by the pulsed field

were detected at the end of a field-free flight tube with a dual microchannel plate detector

in chevron configuration. The ion or electron signal was transferred at each experimental

cycle to a 500 MHz digital oscilloscope (LeCroy, WaveRunner 6050) where they were av-

eraged over a predefined number of cycles. Temporal gates were placed at the position of

the times of flight of the different charged fragments and the signal was integrated over

these gates. The integrated signal was then transferred to a PC and stored on a disc

together with various other signals (laser intensities, calibration signals, etalon signal,

etc.).

2.2 The laser systemsThe experiments were carried using a laser system consisting of a Nd:YAG pump laser

(Continuum, Powerlite Precision 9030) operated at a repetition rate of 25 Hz and two

dye lasers (Lambda-Physik, ScanMate 2E and Sirah, Cobra Stretch). Depending on the

wave length needed for the experiments, either the second (532 nm) or the third (355 nm)

harmonic output of the Nd:YAG laser was used to pump the dye lasers. Moreover, the

frequency of the dye lasers could optionally be doubled using a -Barium-Borate (BBO)

crystal or tripled using two BBO crystals in series, the first to double the fundamental

frequency, the second to generate a beam of a frequency corresponding to the sum of the

frequencies of the fundamental and the frequency-doubled beams.


27/203

The laser systems 11

2.2.1 Experiments using VUV radiation

The investigations of the single-photon photoelectron spectrum of ND3 and ND+3 presented

in Chapter 3 relied on the use of VUV radiation tunable in the range of 80 000 cm1 and

84000cm1. Since no commercial tunable laser systems are available in the VUV region ofthe electromagnetic spectrum, this radiation was generated by resonance-enhanced four-

wave mixing in a pulsed beam of atomic krypton used as nonlinear optical medium [2428].

In these experiments, two additional vacuum chambers, the four-wave mixing chamber

and the monochromator chamber, were used as depicted on the right-hand side of Fig.

2.1. In the four-wave mixing chamber, the nonlinear gas is introduced by means of a

pulsed gas nozzle (General Valve, Series 9, with IOTA One driver; nozzle orifice diameter

of 1 mm).

The VUV radiation generated by the interaction between the focused dye laser radia-tion and the nonlinear gas was then separated from laser beams of different wave lengths

(e.g., the fundamental beams and beams generated by other nonlinear processes) using

a home-built vacuum monochromator of the Seya-Namioka type, equipped with an ad-

justable toroidal platinum-coated diffraction grating and operated in first order. The

toroidal geometry of the grating served the purpose of refocusing the initially divergent

VUV radiation at the exit hole of the monochromator, enabling a clean separation of the

desired beam. The VUV spot size amounted to approximately 1 mm in the photoexci-

taion region. The four-wave mixing chamber, which is exposed to a higher gas load, wasevacuated with a turbomolecular pump of 520 l s1 pumping power (Pfeiffer, TMH 521)

and the monochromator, which is only connected to the other chambers by small holes, by

a 260 l s1 turbomolecular pump (Varian, Turbo V-300 ICE MacroTorr). In this way, the

pressure in the four-wave mixing and the monochromator chambers could be maintained

below 103 mbar and 106 mbar, respectively, during operation of the pulsed valves.

Two laser beams with wave numbers 1 and 2 were used to generate the VUV radiation

by resonance-enhanced difference-frequency mixing in a pulsed atomic beam of krypton.

They were recombined with a dichroic mirror and focused with a spherical lens of 25 cm

focal length at the exit of the pulsed valve delivering the krypton gas beam in the four-

wave mixing chamber. The beam of wave number 1 corresponded to the tripled output

of the first dye laser and was kept at the position of the 4p5 (2P1/2) 5p[1/2]0 4p6 (1S0)

two-photon resonance of krypton (21 = 98 855.0707 cm1 [29]).

Resonance-enhanced sum- or difference-frequency mixing with a second laser beam

of wave number 2 leads to coherent radiation with a wave number VUV = 21 2 asschematically depicted in Fig. 2.3. The VUV wave number was scanned by scanning 2,

thus ensuring resonance enhancement of the four-wave mixing process at all VUV wave

number. The bandwidth of the VUV radiation was typically 0.3 cm1, primarily limited

by the bandwidth of the two dye lasers used for its generation. The VUV intensity was


28/203


monitored using a solar-blind electron multiplier (Hamamatsu). More than 108 photons

per pulse were obtained after the monochromator which turned out to be sufficient to

record PFI-ZEKE photoelectron spectra of ND3 around 82000cm1 (see Chapter 3).

Figure 2.3: Schematic view of VUV generation using resonance-enhanced four-wave mixing in rare

gases.

2.2.2 Experiments using single-photon excitation with UV ra-

diation

The experiments carried out on metastable He and He2 (Chapter 6), ND4, ammoniated

ammonium clusters (Chapter 4) and metastable CO (Section 2.4.3) did not necessitate

VUV radiation. The setup used in these experiments is displayed in Fig. 2.4.

Figure 2.4: Schematic view of the experimental setup used in experiments requiring a single UV laser.

A single dye laser, the output of which was frequency doubled in a BBO crystal, was

employed to excite these species to the region of the ionization thresholds. To avoid


29/203

The laser systems 13

the production of too high concentrations of ions in the photoexcitation volume, the UV

radiation was not focused and its intensity was kept below 100 J/mm2 per 8ns long

pulse.

2.2.3 Experiments using multiphoton excitation

The studies of high Rydberg states of xenon (see Chapter 5) were carried by (2 +

1) resonance-enhanced three-photon excitation with the experimental setup depicted

schematically in Fig. 2.5. The frequency-doubled output of a 355 nm Nd:YAG-pumped

dye laser was employed to induce the 5p5(2P3/2)6p[1/2]0 5p6(1S0) two-photon tran-sition at 21 = 80119.984cm

1 and the high ns and nd Rydberg states located below

the Xe+ 2P3/2 ionization limit were accessed from the 5p5(2P3/2)6p[1/2]0 intermediate

level using the output (wave number 2) of the second dye laser. The two lasers wererecombined using a dichroic mirror and directed in a copropagating arrangement into the

photoexcitation region where they intersected the xenon gas beam at right angles. In these

experiments, care had to be taken to optimize the laser spot sizes in the photoexcitation

region, particularly the spot size of the first laser: Indeed, whereas the efficient driving

of two-photon transitions required a high intensity and some focusing, too tight a focus

led to the production of high concentrations of undesirable Xe+ ions by one-color (2 + 1)

resonance-enhanced three-photon ionization. In these experiments, the nonlinearity of

the multiphoton excitation process and the necessity to work far below the saturationthreshold for the two-photon transition led to unavoidable large signal fluctuations. The

long lifetime of the intermediate state made it possible to delay the second laser pulse

with respect to the first laser pulse using a delay line and thus avoid ac Stark shifts and

broadenings of the transitions from the intermediate level to the Rydberg states.

Figure 2.5: Schematic view of the experimental setup for the xenon experiment. The photon with

wave number 1 was used for the two-photon resonance of 80 118.9839 cm1 in xenon. The photon with

2 was used to prepare the Rydberg state. The millimeter-wave (mm-wave) radiation was used to inducetransitions between two Rydberg states.


30/203


2.2.4 Calibration

Calibration of the dye laser wave numbers was achieved by splitting off small fractions

of their fundamental outputs and directing them towards either an optogalvanic (OG)

cell filled with neon or argon gas or an iodine cell where optogalvanic spectra or laser-induced fluorescence spectra were recorded, respectively. The fundamental frequency was

then determined by comparing the lines in these spectra with tabulated transitions of

neon or argon [30,31] or transitions listed in the I2 atlas [3234]. In addition, a monitor

etalon spectrum of the fundamental output of the scanning laser (laser 2 in Fig. 2.1) was

recorded with every measurement in order to detect and correct possible nonlinearities

in the laser scans. The accuracy of the calibration procedure is estimated to be better

than 0.2 cm1 (OG cell) or 0.05 cm1 (iodine cell) for the fundamental outputs of the dye

lasers. In experiments where VUV radiation was used, the absolute accuracy of the VUVwave number is estimated to be better than 1 cm1, a conservative estimate corresponding

to the sum of the uncertainties of all fundamental laser beams involved in the nonlinear

frequency upconversion process.

2.3 Millimeter-wave source

The high-resolution millimeter-wave spectra of high-n Rydberg states of xenon presented

in Chapter 5 were recorded with the experimental setup depicted in Fig. 2.5 which issimilar to that described in Ref. [35]. The millimeter-wave source we used is a frequency-

and phase-stabilized backward wave oscillator (BWO) operating in the frequency range of

240 - 380 GHz. This millimeter-wave source has been described in detail in Ref. [36] and

the experimental procedure followed to record the millimeter-wave spectra was identical

to that first described in Ref. [35] and subsequently reviewed in Ref. [23]. Only a brief

description of the millimeter-wave source and the measurement procedure is described

here and in Subection 2.5.3, and the reader is referred to the references mentioned above

for a more complete description.

The millimetre-wave source is a heterodyne stabilized synthesizer (see Ref. [37], p.

198). The voltage controlled oscillator is a backward wave oscillator (BWO OB-65, from

ISTOK, Fryazino, Moscow Region, Russian Federation) mounted and aligned in the 30 mm

gap of a 0.7 T Nd-Fe-B alloy permanent magnet (IAPRAS, Nizhnii Novgorod, Russian

Federation). The tube delivers continuous output power up to 40 mW at frequencies

between 240 GHz and 380 GHz. Its operation requires a negative high voltage V0 between

1500 V and 3800 V at an operating current in the range 1035 mA for the cathode andan ac voltage of approximately 6.2 V at 1.4 A for the indirect cathode heating. The heating

voltage is derived from the main power line at a frequency of 50 Hz and is connected to

the cathode potential.


31/203

Probe gas samples 15

The BWO is driven by a commercially available controllable power supply (HCV

375M-6500, F.u.G. Elektronik GmbH, Rosenheim, Germany), which relies on all-solid-

state technology, giving high reliability, a stability of < 5 105 and a repeatability of< 103. Its ripple (peak to peak) is specified to be < (105

|V0

|+ 20 mV), which is to be

expected from a switching topology.

Computer controlled sweeping of the millimeter-wave source is accomplished by vary-

ing the acceleration voltage and local oscillator frequency on the corresponding instru-

ments over GPIB interfaces. The cathode potential is preset according to a tuning curve

fitted to points measured manually:

Vcathode/V = 1387 19.682(f /GHz) + 0.12039(f /GHz)2

2.4599 104(f /GHz)3 + 2.9164 107(f /GHz)4.(2.1)

The effective tuning curve deviates from the calculated one by offsets up to 3 V, whichare compensated for by a phase-lock loop [36].

The frequency stability of this millimeter-wave source and the absolute accuracy of

the frequency calibration are both better than 1 kHz which is negligible compared to the

width of 250 kHz of the narrowest lines observed experimentally so that contributions of

the millimeter-wave source to the uncertainty of the measured millimeter-wave transition

frequencies can be ignored in excellent approximation.

Occasionally, the frequency dependence of the output power of the millimeter-wave

source brought the source out of its lock position, leading to unusable spectra. In order to

be able to identify such problematic regions in the spectra, the lock signal was continuously

monitored during scans of the millimeter-wave spectra.

2.4 Probe gas samples

The study of metastable atoms and molecules, of high Rydberg states and of highly

reactive species such as free radicals and molecular ions by high-resolution spectroscopy

requires the development of adequate methods to produce them. Often, the best method

to generate a given species in a given state is not known a priory and a substantial part of

the experiments is invested in the search of a production method yielding concentrations of

the desired species large enough for spectra of sufficient signal-to-noise ratios (i.e., at least

2) to be obtained during the typical time of four hours of a given measurement session.

Longer measurements times turned out to be impractical because of the unavoidable drifts

of the experimental conditions (stray fields, nozzle operation conditions, laser frequencies,

temperature, etc.).

In the following subsections, the methods used to produce adequate samples for the

spectroscopic investigations of the ammonium radical and ammoniated ammonium clus-


32/203


ters, metastable He and He2, metastable CO and high Rydberg states of atomic xenon

are described as used in the work presented in this thesis. Whereas several production

methods were adapted from descriptions available in the literature as referred to in the

next subsections, the method of producing metastable He2 is original and represents one

of the important results of this dissertation.

2.4.1 Studies of molecular cations using seeded supersonic ex-

pansion

Direct spectroscopic investigations of molecular cations are intrinsically difficult because

space-charge effects limit the concentrations of cations in the gas phase to less than about

108 ions/mm3. Moreover, the total number of cations is usually distributed over many

quantum states because ion samples are typically hot samples. Finally, the high reactivity

of molecular cations often leads to the rapid destruction of the initial sample.

The approach followed by our group to study molecular cations is by pulsed-field-

ionization zero-kinetic-energy (PFI-ZEKE) photoelectron spectroscopy, a technique that

will be described in more detail in Subsection 2.5.2. The advantage of this technique is

that spectra of the cations are obtained from a spectroscopic measurement on the parent

neutral molecule: There are no restrictions on the number density of neutral molecules

imposed by space-charge effects, and, if the neutral molecule is stable, concentrations in

excess of 1014 molecules/cm3 can be obtained in a supersonic expansion. This popula-

tion is distributed over only a very small number of quantum states because of the low

temperatures (in the range of 1-10 K) that can be reached in supersonic expansions (see

above).

The studies of the reactive species ND+3 , ND+4 , ND2ND

+4 , NH3NH

+4 , and He

+2 are

presented in the Chapters 3, 4, and 6. For these ions the experimental task lies in the

preparation of the neutral sample which is trivial in the case of ND +3 because it solely

consists of making an adequate mixture of ND3 and a carrier gas such as argon. In the case

of ND+4 and ammoniated ammonium ions, the task is less trivial because it necessitates

the creation of a suitable neutral ammonia cluster precursor in the supersonic expansion

and a subsequent photolytic reaction to yield the desired radical (see Subsection 2.4.3).

In the case of He2, the task is the most challenging because it necessitates the production

of metastable neutral He2 with more than 20 eV of internal electronic energy and the

formation of a supersonic expansion of these molecules (see Subsection 2.4.2)


33/203


2.4.2 Discharge source of metastable He and He2 in supersonic

expansions

To generate metastable He and He2 (in the following we shall denote these species He

for He 1s2s 3S1 and He2 for He2 a 3+u ), we found it convenient to adapt a pulsed glow

discharge source similar to those described in Refs. [38,39]. Pulsed glow discharges have

distinct advantages compared with continuous glow discharges [4042]. Indeed, because

the voltage and current are applied over short periods of time, higher peak voltages and

currents can be obtained for the same average power, resulting in enhanced sputtering,

excitation and ionization, and therefore higher signal intensities [4346]. Moreover, the

pumping requirements are much reduced.

Figure 2.6: A schematic diagram of the pulsed discharge source.

One of the advantages of the discharge design used to generate metastable He and

He2 is its simplicity. By striking an intense, localized discharge behind a nozzle and by

biasing the electrodes such that electrons flow against or along the gas flow, a stable,

high-current discharge can be produced. In addition, the design allows the aperture to

be independently biased, effectively creating a third electrode in the discharge volume. A

schematic diagram of the discharge source configuration is shown in Fig. 2.6. It consists

of a commercially available pulsed solenoid valve (General Valve, Series 9, with Iota

One driver), several pierced extension pieces and two electrodes made of brass. The

components are electrically isolated and can be biased with respect to one another. The

front plate of the pulsed valve is separated from the cathode by a polyetheretherketone

(PEEK) insulator. The discharge occurs through the channel of a second insulator to the

anode. A high dc voltage is applied to the electrodes through an electrical circuit. The

discharge takes place in a small region between the electrodes, when a high-voltage (HV)

pulse is applied to one of the electrodes just after the gas is pulsed through the valve. The

pulsed valve is operated at stagnation pressures ranging from 1 to 4 bar at a repetition


34/203


rate of 25 Hz for pulse durations between 170 s and 270 s.

To create the discharge, the front electrode is biased positively with respect to the rear

electrode by applying a high-voltage pulse. This arrangement results in electrons being

accelerated in the direction of the gas flow and creates a stable and intense discharge

localized upstream from the exit of the source.

The stability of the operation of the discharge unit can be verified by observing the

reproducibility of the current pulses. In order to deflect charged particles so that they

do not fly through the skimmer into the measurement chamber, a voltage is applied to

deflecting electrodes mounted downstream of the nozzle and connected to the discharge

electrodes.

The operational characteristics of the metastable source are conveniently characterized

by time-of-flight (TOF) mass spectrometry. In the case of He and He2 metastables, the

yield of metastable atoms and molecules can be probed following ionization with a UV

laser by TOF mass spectrometry as illustrated by Fig. 2.7.

TOF spectra such as that displayed in Fig. 2.7 can be used to optimize the production

of a given species by systematically changing the source conditions (nozzle stagnation

pressure, discharge current and voltage, timing of the discharge pulse) so as to maximize

the corresponding signal.

He

+

He2+

-1 0 1 2 3 4 5time of flight /s

0

ionsignal(arb.units)

Figure 2.7: Time-of-flight mass spectrum of He+ and He+2 recorded following photoionization of He

and He2 with a UV laser of wave number 39 603.96 cm1 and extraction of the ions with a pulsed electric

field.

In our attempts to extend this method of generating metastable atoms and molecules

to other rare gas samples and mixtures of rare gases, we found several important limita-


35/203


tions. Firstly, if the He sample is diluted in Ar, Kr or Xe, the formation of metastable

helium atoms and molecules is entirely quenched by a process known as Penning ioniza-

tion, as first described by Penning in 1927 [47]. This process, described by Eq. (2.2),

A + B A + B+ + e (2.2)

occurs very efficiently whenever the internal electronic energy of the metastable state

exceeds the ionization energy of B. The only carrier gas other than helium for which

reaction (2.2) cannot take place in the presence of metastable He for energetic reasons is

neon. The effects of Penning ionization observed after striking a discharge in a mixture of

He and Xe is demonstrated in Fig. 2.8 which shows the distribution of times of flight of the

ions and the metastable species from the discharge region to the detector. The early peak

corresponds to photons produced in the discharge and which reach the detector almostinstantaneously. The next weak peak observed at a time-of-flight of about 40 s can be

attributed to Xe+ ions produced by Penning ionization with metastable helium atoms

and which are accelerated toward the detector. The broad peak centered around 1.1 ms

corresponds to metastable Xe atoms which cannot be destroyed by Penning ionization

and fly as neutral atoms all the way to the microchannel plate (MCP) detector. They are

observed at the detector because their internal energy exceeds the work function of the

front plate of the MCP. No He* atoms are observed.

Xe+

Xe*

0 0.25 0.5 0.75 1 1.25 1.5 1.75time of flight / ms

0

intensity(arb.units.)

Figure 2.8: Time-of-flight of Xe+ after collision and subsequent Penning ionisation with metastable

helium. In this picture the full time-of-flight trace is shown. The peak at t = 0 corresponds to photons

produced in the discharge.

Secondly, of all rare gas dimers, only metastable helium dimers where found to be


36/203


long-lived enough to survive the time of several tens of microseconds needed to travel

from the discharge region to the photoexcitation region. This observation is compatible

with the observations of Oka et al. [48] who found the lifetimes of the triplet states to

decrease rapidly from He2 to Xe2 because of the growing admixture of singlet (S = 0)

character induced by the increasing spin-orbit interaction.

2.4.3 Optical preparation of metastable states

Metastable states of atoms and molecules can also be prepared by driving forbidden

transitions with intense laser fields. Two cases have been examined in this thesis: the

production of metastable high Rydberg states of Xe by (2+1) resonance-enhanced three-

photon excitation for subsequent investigation of millimeter-wave transitions between high

Rydberg states of Xe (see Chapter 5) and the preparation of the metastable lowest tripletstate (a 3) of CO by single-photon excitation from the X 1+ ground state.

(a)

Laser

CO*

0

(b)

CO

+

C++ O

+

CO*

0intensity(

arb.units)

(c)

0 100 200 300 400 500time of flight /s

0

Figure 2.9: Time-of-flight spectrum illustrating the production of metastable CO

a 3=1 moleculesin a supersonic expansion. a) TOF spectrum recorded without applying any voltage on the extraction

plates surrounding the photoexcitation region. b) TOF spectrum recorded with a voltage of 1 kV on the

repeller plates of the extraction region. c) TOF spectrum recorded under the same conditions as b) but

after detuning the UV laser.

The formation of metastable CO (a 3=02, denoted below as CO) following single-

photon excitation from the CO X 1+ ground state at a wave number of 48 486 cm1 is

illustrated by the TOF spectra depicted in Fig. 2.9. The top trace (a) was obtained by

monitoring the CO molecules arriving at the MCP detector where their internal energy

(6 eV) causes ionization because it exceeds the work function of the alloy at the MCP


37/203


surface. The middle trace (b) was obtained under the same conditions as trace (a) with

the only difference that a voltage was applied to the extraction plates to also accelerate

ions produced by the UV laser. The ions (C+, O+ and CO+) are observed as sharp

features of times of flight of

10 s. The broad peak corresponding to CO is unaffected

by the voltage which proves that the metastables fly as long-lived neutral molecules all

the way to the detector for 400 s. The bottom trace (c) was recorded under the same

conditions as trace (b) but after detuning the UV laser. This trace demonstrates that

CO is produced by optical excitation from the ground state.

Spectra of the forbidden a 3=02 X 1+ transition can be recorded by settingan integrating time gate at the position in the TOF spectra corresponding to the CO

molecules and monitoring the integrated signal as a function of the wave number of the

UV laser. An example illustrating the production of CO a 3=1 metastables is depicted

in Fig. 2.10. The spectral features observed in this spectrum correspond to the formation

of low rotational levels of the a 3=1 states from the lowest (J = 0, 1 and 2) rotational

levels of the X 1+ ground state of CO. The signal-to-noise ratio in this spectrum is

limited by the large width of the integrating time gate and fluctuations of the UV laser

intensity.

CO a 31 CO X1+

2 2

1 1

0 1

1 2

2 3

48460 48470 48480 48490 48500

wave number / cm-1

0

intensity(arb.units)

Figure 2.10: Spectrum of the forbidden a 3=1 (v = 0) X 1+ (v = 0) transition of CO

obtained by recording the CO molecules at the MCP detector after a flight time of 400 s as a function

of the wave number of a UV laser.

The possibility of generating metastable a 3 CO molecules is not only interesting

in view of high-resolution spectroscopic studies of high Rydberg states of CO having

S = 1 character but also in the context of recent experiments in our group in which


38/203


multistage Zeeman deceleration and magnetic trapping is used to generate cold molecular

samples [49,50].

2.4.4 Photolysis source of metastable radicals in supersonic ex-

pansions

Recently, Willitsch et al. [51] have developed a source of cold radicals in supersonic ex-

pansions that relies on the photodissociation of suitable precursor molecules in a quartz

capillary mounted at the orifice of a pulsed valve. The formation of a supersonic ex-

pansion into the vacuum at the exit of the capillary led to very efficient cooling of the

internal (rotational and vibrational) degrees of freedom of the radical. In combination

with PFI-ZEKE photoelectron spectroscopy, this method turned out to be a powerfulway to obtain spectroscopic informations on the neutral radicals and the cations formed

by photoionization, as illustrated by studies of the PFI-ZEKE photoelectron spectrum of

NH2 [52], C5H5 [53], and CH3 [54].

Unfortunately, when the radicals are unstable and decay by predissociation on the mi-

crosecond time scale or faster this method cannot be employed because the radicals decay

during the flight from the exit of the capillary orifice to the region where the photoioniza-

tion is carried out. In this case, it is more convenient to carry out the photodissociation of

the precursor molecule directly in the photoionization region. An example of such an un-stable radical is the ammonium radical studied by PFI-ZEKE photoelectron spectroscopy

by Palm et al. [55] and Signorell et al. [56] (see also Subsection 4.3.1). The study of

metastable ammoniated ammonium radicals presented in Chapter 4 relied on the same

method as used in Refs. [55, 56]. The optimization of the radical cluster yield was car-

ried out by photoionization mass spectrometry illustrated in Fig. 2.11 for ammoniated

ammonium radical clusters formed by photodissociation of ammonia clusters.

The top trace in this figure displays the TOF spectrum obtained by collecting the ions

produced by the 206 nm photolysis laser only, i.e., the laser producing the ammoniated

ammonium clusters from ammonia clusters (The production mechanism will be discussed

in more detail in Chapter 4). The middle trace (b) in Fig. 2.11 shows the TOF spectrum

obtained after photoexcitation/photoionization with a second UV laser of wave number

32 051.28 cm1 delayed by 15 ns with respect to the photolysis laser, and the bottom

trace (c) shows the same spectrum from which, however, the ionization signal from the

photodissociation laser was subtracted. This trace is free from background ionization

signals from the photodissociation laser and enables one 1) to determine and optimize

the yield of metastable radicals produced by the photodissociation laser, and 2) to study

these radicals without interfering contributions to the signals from the photodissociation

laser.


39/203

Spectroscopic methods 23

(a) TOF whithout photoionization laser

NH3

+NH

4

+(NH

3)2

+NH

3NH

4

+(NH

3)2NH

4

+

0

(b) TOF at 32051.28 cm-1

(NH3)3NH

4

+

(NH3)4NH

4

+

0ionsignal(arb.units)

(c) difference spectrum

5 6 7 8 9time of flight /s

0

Figure 2.11: Ion TOF mass spectra of ammonia and ammoniated ammonium radical clusters. a)

Background TOF mass spectrum generated by the photodissociation laser alone. b) TOF mass spectrum

including background signal, obtained following photodissociation and photoionization. c) Difference

spectrum obtained by subtracting trace a) from trace b). This spectrum enables one to observe the

photoionization of the photodissociation products with interference from undesirable photoionization by

the photodissociation laser.

2.5 Spectroscopic methods

All experiments carried out in the realm of this dissertation were performed by detect-

ing charged particles (electrons or cations) produced either by photoionization or by the

pulsed-electric-field ionization of long-lived Rydberg states. Detecting charged particles

is very efficient efficiencies of 50% or more are routinely achievable and offers the

important advantage of being background-free, i.e., the signal vanishes outside the reso-

nances. These advantages result in a very high sensitivity which represents an essential

advantage in studies of reactive or unstable species that are only formed in low concen-trations. They are at the heart of the three spectroscopic methods used in this thesis and

which are described in the next subsections: photoionization spectroscopy in Subsection

2.5.1, pulsed-field-ionization zero-kinetic-energy (PFI-ZEKE) photoelectron spectroscopy

in Subsection 2.5.2 and millimeter-wave spectroscopy of high Rydberg states in Subsection

2.5.3.

2.5.1 Photoionization spectroscopy

In photoionization spectroscopy, the yield of charged fragments produced following pho-

toexcitation of a molecule with a tunable light source is monitored as a function of the


40/203


frequency of the light source [57,58]. In the most commonly used variant of photoioniza-

tion spectroscopy, the fragment ions produced after photoexcitation of the parent molecule

(designated here by ABC) are monitored by time-of-flight mass spectrometry which en-

ables one to distinguish between different product channels:

ABCh ABC+ + e (2.3)

ABCh AB + C+ + e (2.4)

ABCh A+ + BC + e (2.5)

ABCh AC + B+ + e (2.6)

ABCh AB+ + C + e (2.7)

ABCh

A + BC+ + e (2.8)

...

Process (2.3) corresponds to the photoionization of the parent molecule, and processes

(2.4) through (2.8) describe several possible dissociative ionization channels. Several con-

tinua corresponding to the energetically accessible quantum states of the charged frag-

ments can contribute to the ion signal recorded at a given mass and, usually, the ion

yield displays resonances corresponding to photoexcitation to Rydberg states of the par-

ent molecule ABC that belong to series converging on excited states of the cation ABC +,

as illustrated in Fig. 2.12 and described in its caption. Photoionization spectra are there-

fore highly structured, very complex and difficult to analyze. Usually, photoionization

spectra are used to determine the appearance thresholds of the fragments, e.g. ABC+,

AB+, A+, etc., which, in favorable cases, can be further used to derive thermochemical

quantities such as the adiabatic ionization energy of ABC or the dissociation energies

of ABC+ in various fragments such as AB+ + C, A+ + BC, etc. In general, the pres-

ence of autoionization resonances prevents the unambiguous determination of ionization

thresholds corresponding to excited states of the fragments, and renders the appearance

of photoionization spectra extremely complex [16,57,58]. The photoionization spectra of

He2, presented in Chapter 6, represents an excellent example of this complexity (see, e.g.,

Fig 6.8).

To record a photoionization spectrum, the ions are usually extracted from the pho-

toionization region by an electric field and directed through a field-free flight tube towards

a detector, typically a microchannel plate detector. The electric field also has the effect of

lowering the ionization thresholds and this effect must be corrected in the determination

of ionization energies. When a dc field is used to extract the ions, the lowering of the

ionization thresholds can be accurately determined using a simple classical picture by

calculating the saddle point in the potential V(r) resulting from the sum of the Coulomb


41/203


Figure 2.12: Energy level diagram illustrating the different photoionization channels of a polyatomic

molecule. The ionization and dissociative ionization thresholds are marked by thick horizontal bars above

which the shaded areas represent continua. These thresholds are characterized by the quantum states of

the fragments. Infinite series of Rydberg states converge to each threshold. The full arrow designates

direct ionization; the thick blue arrows represent ionization processes via excited states of the parent

molecule, giving rise to so-called autoionization resonances in the photoionization yield; the dashed (red)

arrow represents the process of dissociative ionization.

potential binding the electron to the cation VC(r) = e2

40r , where r = |r| and r = (x,y,z)represents the position vector of the electron relative to the center of charge of the cation,

and the potential of the electron in the applied electric field (eEz, if the field is applied inthe z direction, i.e., if E= (0, 0, E)):

V(r) = e2

40r+ eEz. (2.9)

This saddle point is shifted from the field-free ionization energy EI to lower energies by

EIhc cm1

= 6.12

E

V / c m(2.10)


42/203


as can be easily determined from Eq. (2.9).

When a pulsed field is used to extract the charged fragments, the details of the field

ionization dynamics must be considered in the determination of the field-induced lowering

of the thresholds [1416, 59]. In general, a reasonable approximation for the lowering of

the ionization thresholds when the pulsed field does not exceed 10 V cm1 and is applied

with a rise time in the ns range is [15,16]

EIhc cm1

= 4

E

V / c m. (2.11)

If the applied field is larger and applied with slower rise times, Eq. (2.10) represents a

better approximation than Eq. (2.11).

Photoionization spectra can be recorded either directly from the ground state, or froma metastable state following single-photon excitation, or via an intermediate excited state

of the parent molecule using several photons. In the latter case, one usually calls the pho-

toionization spectra REMPI spectra for resonance-enhanced multiphoton ionization

spectra.

2.5.2 PFI-ZEKE photoelectron spectroscopy

Photoionization spectroscopy offers the advantage of being mass-selective but usuallycannot be used to determine any ionization threshold other than the lowest one because

of autoionization processes. Photoelectron spectroscopy is exactly complementary. It

enables one to determine the ionization thresholds by measuring the kinetic energy of the

photoelectrons but it does not provide any information on the mass of the parent molecule

or the ionic products.

Photoelectron spectroscopy has a long history in studies of the electronic structure

of atoms and molecules [60, 61]. In the present dissertation, a variant of photoelectron

spectroscopy called pulsed-field-ionization zero-kinetic-energy (PFI-ZEKE) photoelectron

spectroscopy [6,7] has been used.

In a PFI-ZEKE photoelectron spectrum, one monitors the yield of electrons produced

by the delayed pulsed-field-ionization of very high Rydberg states located just below the

ionization thresholds as a function of the frequency of the photoionizing radiation (see Fig.

2.13b). This is in contrast to conventional (He I) photoelectron spectroscopy which relies

on the measurement of the kinetic energy distribution of the photoelectrons produced

following photoexcitation with radiation of fixed frequency (see Fig. 2.13a).

The advantage of PFI-ZEKE photoelectron spectroscopy over He I photoelectron spec-

troscopy results from the fact that the measurement of the electron kinetic energies, which

is difficult, is bypassed (see Ref. [7] for details).


43/203


Figure 2.13: Comparison of conventional photoelectron spectroscopy (panel (a)) and PFI-ZEKE pho-

toelectron spectroscopy (panel (b)). With the former technique the ionization energies EI are determined

by ionizing with a fixed-frequency light source (typically a He I source) and measuring the kinetic energy

of Ekin of the photoelectrons using EI = h Ekin. With the latter technique, they are determined bydetecting the field-ionization of high Rydberg states as a function of the frequency of a tunable light

source.

The electrons contributing to a PFI-ZEKE photoelectron spectrum stem exclusively

from the field-ionization of long-lived high Rydberg states because electrons emitted in

the ionization continua escape from the photoexcitation region before the pulsed electric

field is applied. A PFI-ZEKE photoelectron spectra thus provides a map of the relative

positions of the ionic energy levels with respect to those of the parent neutral molecule.

Usually, small pulsed electric fields are used so that Eq. (2.11) can be used to deter-

mine the field-free ionization thresholds from the positions of the lines in the PFI-ZEKE

photoelectron spectra.

The best resolution achievable by PFI-ZEKE photoelectron spectroscopy is obtained

using multipulse electric field sequences [59, 62, 63]. In combination with narrow-band

tunable VUV laser light sources, multipulse electric field sequences enable the recording

of phototelectron spectra at a resolution of better than 0.1 cm1 [59] which is sufficient

to observe the full rotational structure in the spectra of polyatomic molecules such as

ethylene [64], ozone [65], the cyclopentadienyl radical [66] and allene [67].

The line intensities in PFI-ZEKE photoelctron spectra do not only reflect the direct

photoionization cross sections but are usually influenced by autoionization resonances

of the kind depicted in Fig. 2.12 (see Refs. [16,68] for more details). Examples will be


44/203


discussed in the Chapters 3 and 6. The analysis of the intensity distributions in PFI-ZEKE

photoelectron spectra thus provide valuable information on the interactions between the

different ionization channels [69,70] and can be carried out using multichannel quantum

defect theory (MQDT) [71,72].

2.5.3 Millimeter-wave spectroscopy of high Rydberg states

By measuring transitions between neighboring high Rydberg states using millimeter waves,

it is also possible to obtain spectroscopic informations on high Rydberg states and, by

extrapolation of the Rydberg series, to determine the ionization thresholds [23, 73, 74].

In comparison to PFI-ZEKE photoelectron spectroscopy, a much higher resolution and a

much higher precision can be achieved in the determination of ionization thresholds for

several reasons: Firstly, millimeter-wave radiation sources have bandwidths of less than1 kHz. Secondly, the Doppler width of spectral lines, which is proportional to the fre-

quency, is extremely small in the millimeter-wave range of the electromagnetic spectrum.

Thirdly, the continuous-wave nature of the millimeter-waves permits very long measure-

ment times, which translates into a very h

Documents

Matthias Raunhardt- Generation and spectroscopy of atoms and molecules in metastable states